V1 RESPONSES TO TRANSPARENT AND NONTRANSPARENT MOTIONS

It is well known that a stimulus composed of two independent sets of r andom dots moving in opposite directions produces a percept of two ove rlapping transparent surfaces moving across each other, while a counte rphase grating composed of two identical sine wave gratings drifting i n opposite directions does not. We recorded from the directionally sel ective V1 cells of behaving macaque monkeys using these two types of s timuli in order to investigate the physiological basis of transparent motion perception. Previous single-unit recording experiments from our laboratory indicated that many V1 cells respond well to transparent r andom dot patterns, while MT cells' responses to the same patterns are strongly suppressed in comparison with their preferred direction resp onses. This observation alone would seem to suggest that V1 activity c ould better explain transparent motion perception than MT activity. Ho wever, one could argue to the contrary based on the psychophysical obs ervation that there is a motion threshold elevation under the transpar ency condition. We decided to determine the correlation between V1 act ivity and the transparent motion perception directly by recording from V1 cells using both transparent random dot patterns and nontransparen t counterphase gratings. It is found that V1 cells on the average coul d not reliably tell the two types of patterns apart. Our results furth er the idea that additional processing beyond Vf is involved in transp arent motion analysis.